July, 1945
THESEROLOGICAL PROPERTIES OF
For malachite green (Fig. 11) no influence of varied concentrationand only an extremely small, perhaps doubtful one, in the presence of agar, has been observed. In phenosafrculin (Fig. 12) no influence of concentratiop in aqueous solution can be seen, In agar the curve is slightly depressed but not displaced. The behavior of the “normal” color in nucleic acid is the same as for strongly metachromatic dyes. Figure 13 shows the ultraviolet band of methylene blue is independent of concentration and of agar, within the limits of error. Figure 14 shows that also in thionine the ultraviolet band depends very little on the concentration,if at all, and even on extreme variation of concentration neither the peak of absorption is displaced nor any secondary band is established.
1219
SIMPLE SUBSTANCES
mp. 400
350
300
275
I
1
250 I
225 1
50,000
130,000
summary Many basic dyestuffs are adsorbed by stainable substrates in two different shades of color, designated as the normal and the metachromatic color. As a model for a substrate stainable in the normal color, a 3y0 solution of nucleic acid a t p H 4.6, is chosen; as a model for a substrate stainable in the metachromatic color a solution or gel of agar a t PH 4.6 is chosen. These model substrates allow of spectrophotometric measurement of the absorption curve. It is shown that all dyestuffs capable of metachromasy disobey Beer’s law in aqueous solution, due to the fact that with increasing concentration dimeric molecular aggregates of the dye molecules are formed. The absorption maximum of the dimer lies a t shorter wave lengths than that of the monomer as it exists in extremely dilute solution. In the presence of agar still higher molecular aggregates are formed and it is these aggregates which are adsorbed by agar. The absorption bands of these high polymers are still further displaced toward shorter wave lengths and are more diffuse. In contrast, in the presence of nucleic acid the absorption
10,000
22,000
26
30 34 v(cm.-l). Fig. 14.
38
42,000
spectrum of all basic dyes is independent of the concentration of the dye and is similar to, although not identical with, that of the dye in extremely dilute solution. No polymerization takes place. Each cation of the dye is combined with one acidic side chain of the nucleic acid to form a stoichiometrically well-defined salt-like compound. Very little is known about the correlation of the chemical structure of a dye with its faculty of polymerization. On the other hand, this faculty of polymerization is always correlated with the metachromatic effect. NEWYORK,N. Y.
RECEIVED MARCH 23, 1945
FROM THE GATESAND CRELLIN LABORATORIES OF CHEMISTRY, CALIFORNIA [CONTRIBUTION , INSTITUTE OR TECHNOLOGY No.10031
The Serological Properties of Simple Substances. X. A Hapten Inhibition Experiment Substantiating the Intrinsic Molecular Asymmetry of Antibodies BY DAVIDPRESSMAN,
JOHN
H. BRYDEN, AND LINUSPAUZING
It was shown by Landsteiner and van der Scheerl that by inoculating animals with suitable antigens antisera can be produced which can distinguish between optical isomers. Their first experiments were carried out with antisera prepared with use of immunizing antigens made by coupling proteins with diazotized d- and l-p-aminobenzoylphenylaminoacetic acid ; each of the two antisera precipitated preferentially the test azoprotein containing the corresponding haptenic group, and the precipitation was inhibited preferentially by the corresponding hapten. Similar results were also obtained with azoproteins con(1) K . Landsteiner and J. van der Scheer, J . E x p f l . M c d . , 48, 315 (1928); 60, 407 (1929).
taining azo derivatives of the stereoisomeric tartranilic acids as haptenic groups. These experimental results show the significance of spatial configuration in serological reactions. They do not, however, depend in any way on the fact that antibodies themselves have inherent optical activity. Because of the optical activity (2-configuration) of the amino acid residues of proteins, the possibility exists that an antiserum made by use of an antigen prepared from an optically inactive haptenic substance may combine preferentially with one of a pair of optically isomeric substances. We have prepared such an antiserum (anti-& serum), by injecting rabbits with an azoprotein made from the inactive
1220
I)AVID
PRESSMAN, JOHN H. BRYDEN AND LINUSPAULING
VOl. 67
( a , + 3 i . 2 " , 1 dm.); I-, 0 3 9 . 4 " ( a , - 3 i . 3 " , I dni.); reported' [alUD'd-, +40.16 The differences in the rotations of the salts (5:'6) and acids (2%) are undoubtedly due to the presence of the optical isomer. The presence of a non-optically active substance is doubtful since the method of preparation of dl-a-methylbenzylaniine from formamide and acetopheExperimental Methods none makes very iniprobable the presence of any other Protein Antiguur.-The immunizing antigen was pre- amine and the result of the equivalent weight determinapared by diazotizing 0.006 mole of p-aminosuccinanilic tions shows that succinic acid is not present. p-NitrosuccinPnilic acid was prepared by Dr. W. B. Renacid and coupling with 100 ml. of sheep serum at PH 9. frow by mixing hot solutions of 0.2mole of succinic anThe antigen was purified according to the directions of hydride and 0.2 mole of p-nitroaniline in 5O-ni1. portions Landsteiner and van der Scheer.1 The test antigen (S,-ovalbumin) was prepared by di- of dioxane and heating in a boiling water-bath for five azotizing 0.050g. of p-aminosuccinanilic acid and coupling niinutes. The P-iiitrosucciriaiiilic acid crystallized on the product with 0.50 g. of ovalbumin a t about pH 9. cooling; in. p. 194-195', reported 194 -195'.6 pAmino6uccjnanilic acid was prepared by Dr. W. B. The azoprotein was dialyzed overnight against tap water, precipitated twice at PH 3.5 froin 50 ml. of solution, and Kenfrow by reducing 0.08 mole of p-nitrosuccinanilic acid i n 175 ml. of 90yi methanol with hydrogen a t 3 atm. over finally dissolved in saline solution a t PH 7. Antiaera.-The method of preparing and pooling anti- a catalyst of palladium on calcium carbonate. The sera was similar to that described previously for the aminosuccinanilic acid precipitated on cooling. It was dissolved in the theoretical amount of sodium hydroxide preparation of anti-R, sera.' solution, filtered, and reprecipitated with the theoretical Reacti~mof Antiserum with Antigen urd Hapten.-The reactants were mixed and permitted to stand ',or one hour amount of hydrochloric acid; n i , p. 184-185",reported 183-1234 O . 6 a t room temperature and over two nights at 5 . The precipitates were centrifuged and washed three times with Discussion 10-ml. portions of 0.9R sodium chloride solution and were The results of the hapten inhibition experianalyzed by our standard methodVa ments are given in Taple I and Fig. 1. The Preparation of Substances values 0.8 to 3.5 of the heterogeneity index u obSuccbmilic a d d was prepared by the method of tained on application of the theory of heteroAuwers4 by slowly adding 0.16 mole of aniline to a boiling geneous antisera7 to the data are similar in magnisolution of 0.15mole of succinic anhydride in 150 ml. of chloroform. The reaction took place immediately with tude and in trend to those found for other sysprecipitation of succinanilic acid. The product was re- tems. The values of the hapten inhibition concrysta$eed from water, with a yield of 68(;1,,m. p., 147.2- stant KO'(aside from the difference for the d and 147.7 . reported 148.5°.4 Acidic equivalent rueigkt: 1 haptens) are reasonable; succinanilic acid, which calcd. ior C&IO&, 193.1; found, 191.4,191.7. d- and ~N-(armetbylb~l)nzyl)ucciarmio a d d s were pre- is very closely related in structure to the haptenic pared similarly using approximately 0.08 mole of d- or group of the immunizing antigen (the N-(+-azol-a-methylbenzylamine with an equivalent amount of phenyl)-succinamic acid group), combines five to succinic anhydride. It was necessary to extract the product from the chloroform solution with sodium hy- ten. times more strongly with the antibody than droxide. Practically the theoretical quantity was used. the optically active haptens, and about fifty The substituted succiiiamic acid was precipitated with times more strongly than the simple hapten hydrochloric acid and recrystallized from wfter. Yields succinic acid. were :bout f357$; ni. p.: d-, 100.9-101.1 ; I - , 100.9Data were obtained for three different amounts 101.2 , [ C X ] ~ofD sodium salt a t PH 9 in water: d-A of the precipitating antigen, &-ovalbumin, ex+96.7" (a,+2.14", 2 dm., 11.09 g./liter); I-, -91.4 (a,-2.01°, 2 dm.,11.01 &/liter). [a]%D of free acid in tending into the regions of antibody excess and water: d-, +129.2" (a, +2.59', 2 dm., 10.01 g./liter); 1-, antigen excess. It is interesting that in these ex127.0"( a , -2.54", 2 dm., 10.00g./liter). Acidic epuiuulent weight: calcd. for C12HI~OaN,221.1; found, d-, 220.9, periments a given amount of hapten was found
substance p-aminosuccinanilic acid, and have found by hapten-inhibition experiments that the antibodies in the serum combine more strongly with ,!-N-( a-methylbenzy1)-succinamic acid than with the d isomer.
.
-
221.5; I - , 221.6,221.9.
The d- and 1-a-methylbeiizylamines were prepared from redistilled dl-a-methylbenzylamine (b. p. 186.0-186.R" urlcor.) by the method of Lovh.' The dl-amine (0.X34 mole) was added to 0.436 mole of 1-malic acid in &%5 mi. of water. d-a-MethylbenzylamnioniuniI-malate crystallized out overnight with a crude yield of 9.1". The mother liquor was made basic with sodium hydroxide and the amine extracted with ether. The amine was recovered by evaporation of the ether and was added to 0.22mole of d-tartaric acid in 130 ml. of water. I-a-Methylbenzylammonium d-tartrate crystallized overnight with a crude yield of 61yo. The crude salts were crystallized twice from 100 ml. of water. The d- and l-amines were recovered by liberation of the free amines by sodium hydroxide solution, extraction with ether, and subsequent fractional distillation. Yield was d-, 44'1;; 1-, 3870; b. p.: d-, 185.5-186.0' uncor.; 1-, 186.0-186.5' uncor.; D.: d-, 0.947516~s~, I-, 0.94f3W6r; [U]'~.~D:d-. +39.2 (2) L. Pauling, D.Pressnun, D. H. Campbell, C. Ikeda, and M. Ikawa, TmB JOURNAL, 64,2994 (1942h (3) D. Pressman, I n d . Enp. Chrm., A d . E d . , 16, 357 (1943). (1) K,Auwero, A n n . , 809, 320 f1894). (,-!)
J . h l . I.ovCn, . I . p i ' ( i k t . ( . h r r ~ r, 73, 310 tl!lO.->j
to be more effective in inhibiting precipitation the greater the amount of antigen; this corresponds to the simple theory of homogeneous antibody,x but is contrary to earlier work with another type of antiserum.' The striking result of the experiments is thc difference in inhibiting power shown by the il and I isomers of N-( a-methylbenzy1)-succinamicacid, amounting in the region of antibody excess to a factor in Ki greater than 2, and to a difference of about 500 cal. pet mole in the standard free energy of combination of hapten and antibody. This difference is presumably to be attributed to the presence of optically active amino acid residues in the antibody yglobulin, causing the stable con( 6 ) K . Landsteiner and J . van der Scheer J . Ex#. Med., 66, 399 (1932). (7) L. Pauling, U. Pressmiin, and A . 1.. Grossberg, T H I S JOURNAL, 66,784 (1944). ( 8 ) I.. Patiling, I ) I ' r c s t n a n . 1) H.C n m p l ~ e l land . C Ikeila. i h i d 64, X l W i I I!l421.
THESEROLOGICAL PROPERTIES OF
July, 1945
1221
SIMPLE SUBSTANCES
TABLEI EFFECT OF
HAPTENS ON
THE PRECIPITATION OF
ANTI-S, SERUM WITH SI-OVALBUMIN
Antigen solution in borate buffer at PH 8, 1 ml.; antiserum, 1 ml.; hapten solution in saline, 1 ml.; pH of supernates 8.2." Amount of antigen, cg
Hapten,
36
d-I 1-1 d-I
110
1-1
K: 0.070 .16i ,092 .166 1 .w
15.6
31.3
U
Moles of hnpten nddcd, X 101 62.5 12p 250 Amount of precipitnte'
845
3.5 a .5 3.5 2.5 L.5
908 927 785
%4 864 852
724 777 (791) 439
720 594 717 685
500
1000
657 540 642 553
577 435 485 397
I1 883 111 864 799 440 d-I 0.17 3.0 903 (834) (749) 65ti 552 394 543 363 1-1 .19 2.0 887 (871) 7% 702 I1 1 .oo 0.8 966 (877) 74 1 503 I11 927 882 I = N-(a-Methylbenzy1)-succinamicacid; I1 = succinanilic Optimum precipitation 850 pg. at 220 pg. of antigen. acid; 111 succinic acid. a The amounts of precipitate are in parts per mille of the amounts in the absence of hapten: 1180,660, and 239 pg. for 440,110 and 35 pg. of antigen, respectively. Blanks of serum and buffer 8, 8, and 16 pg., respectively. Values are averages of triplicate analyses, With mean deviation *2%, except for duplicate analyses in parentheses.
-
figurations complementary to the symmetric more strongly with the I isomer than with the d haptenic group of the immunizing antigen to be isomer, and that these preferential effects would asymmetric. An alternative but less likely ex- largely cancel each other. A very good antibody planation could be based on the presence of op- molecule, however, would have to assume a very tically active amino acid residues in the protein well-defined configuration, bringing it into the part of the immunizing antigen: this explanation closest approximation to the haptenic group and is rendered improbable by the fact that the part locating and orienting its charged groups and of the haptenic group in the immunizing antigen hydrogen-bond-forming groups in the most satiswhich corresponds to the asymmetric carbon atom of the d and I haptens is separated fram the asymmetric car- . bon atom of the amino acid residue by 4 a considerable distance (the length of ~3 a benzene ring, azo group, and the '& amino acid side chain). b. Our present knowledge of the struc- "0 ture of antibodies is not sufficiently 2 0.5 detailed to permit an interpretation 8 to be made of the observation that it 2 is the I isomer which combines the more strongly with the antibody. MOLWT Cf ANTIGEN A reasonable explanation can be 3 given of the observed dependence 35 PO 110Pg 4 4 0 Po of the optical isomer effect on the 0.0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 amount of precipitating antigen and on the amount of hapten. The antiLogarithm of amortiit of hapten. Semm 's heterogeneous-it contains Fig. 1. Effect of d- and I-N-(a-niethylbeiizy1)-succinarriicacid irr iuantibodies which in configuration ap- hibiting the precipitation of anti& serum arid S,-ovalbumin; points for proximate Only to the hap- the d and 1 haptens, respectively, are represented by the right and left tenic group of the immunizing azo- halves of circles. Drotein and the DreciDitatinn azoDroiein, and hence'exek onlyYweak attraction for factory way for attracting and holding the hapthese groups, and antibodies which fit very closely tenic group! the nature of this most completo the haptenic group, and form strong bonds mentary configuration would be determined by with it. It is to be expected that weak anti- the fact that the building stones are I-amino acid bodies, fitting the haptenic group only loosely, residues, and the configuration would accordingly could be built in a great many different ways, be expected to correspond more closely to one thaii of which some would combine more strongly with to the other of an enantiomorphic pair of asytnthe d isomer of a pair of asymmetric haptens rnetric molecules. th:in with the 1 isomer, ;uid others would cctinbiiic When a sni;ill iitiiount of aritigrii is added to
-8
-
NOTES
1222
portion of antiserum it is the antibody molecules with the greatest attraction for the antigen which form the precipitate with it. The argument given above leads to the expectation that the optical isomer effect should be larger in this case than when a large amount of precipitating antigen is used. This expectation is confirmed by the experimental results. Moreover, in the region of the equivalence zone, with the precipitate containing poor antibody as well as good antibody, the part of the precipitate which is dissolved first on addition of hapten is that containing the poor antibody (with small combining power for the haptenic group of the immunizing antigen and precipitating antigen) ; the good antibody tends to remain in the undissolved precipitate. Accordingly, by the argument given above, the difference in effect of d and I haptens would be small a t low hapten concentrations and larger a t high hapten concentrations, at which a considerable fraction of the precipitate is dissolved. This effect, shown clearly by the middle pair of curves in Fig. 1, leads to a difference in apparent heterogeneity of the antiserum with respect to the isomers. The small difference shown by the isomeric
Vol. 67
haptens in the region of slight antigen excess may be the result of the action of the excess of antigen in favoring the formation of soluble complexes involving good antibody molecules. This investigation was carried out with the aid of a grant from The Rockefeller Foundation. We wish to thank Dr. W. B. Renfrow, Jr., and Mr. Dan Rice for assistance in analyses and the preparation of compounds. Summary An antiserum has been prepared by injecting rabbits with an azoprotein made by coupling sheep serum with diazotized p-aminosuccinanilic acid, the molecules of which have a plane of symmetry. It has been found that the d and 1 isomers of N-( a-methylbenzy1)-succinamic acid differ in their power to inhibit the precipitation of this antiserum by an azoprotein made by coupling ovalbumin with diazotized p-aminosuccinanilic acid, the 1 isomer having the greater inhibiting power. This behavior is presumably an effect of the presence of optically active amino acid residues in the antibody molecules. PASADENA 4, CALIFORNIA RECEIVED APRIL7, 1945
NOTES A New Synthesis of 2-Diethylaminoethyl p-Aminothiolbenzoate
ported.6 Both of these attempts were based upon the reaction of sodium hydrosulfide with 2-bromotriethylamine hydrobromide. In the present BY NOELF. ALBERTSON AND R. 0. CLINTON work we have been able to prepare the desired 2-Diethylaminoethyl p-aminothiolbenzoate compound in about 40% yield by the reaction of (T'hiocaine) was first prepared by Hansen and alcoholic sodium hydrosulfide with 2-chlorotriFosdick,' and was shown to have anesthetic ac- ethylamine. A more satisfactory procedure, tivity. Later work2 indicated this activity to be however, involves the reaction of 2-chlorotrifrom four to six times that of procaine hydro- ethylamine or 2-chlorotriethylamine hydrochlochloride, while the toxicity was not proportion- ride with thiourea, followed by alkaline hydrolysis ately increased. The compound in the form of its of the isothiouronium salt. The amine is less hydrochloride was, however, relatively unstable satisfactory than the hydrochloride, since the in ~ o l u t i o n .These ~ facts suggested that a further tertiary amino group is sufficiently basic to produce partial decomposition of the isothiouronium investigation of its salts was desirable. The simplest synthesis of this substance would chloride with resultant loss of product. Hansen and Fosdickl have reported a melting involve reaction between 2-diethylaminoethanethiol and p-nitrobenzoyl chloride, followed by point of 52.0-52.5' for their base. In the present reduction of the thiol ester. An attempt to use work material of this melting point was obtained this method was made by Lischer and J ~ r d a n , ~in one experiment; however, all other preparabut they were unable to prepare the thiol. A tions gave a base with melting point 75.0-75.5'. subsequent failure to prepare it also has been re- Seeding a melt of the low melting form with the high melting form produced a conversion to the (1) Hansen and Fo;dick, TIixs J O U R N A L , 6G, 2872 (1033), U. s. latter. It is therefore evident that the comPatent 2,090,756. pound exists in dimorphic forms. (2) Fosdick and Hnnaen. J . P A u r m a c u l . , 60, 323 (1934): h-olle, F o r m . i. Farrnakul. ( E S . 5. R.1, 1937,No. 2 I ; ICIxm. A b . > t r . , 34, A number of new salts of 2-diethylaminoethyl 3820 (1940)). p-aminothiolbenzoate were prepared. Of these ('3) Private communication f r o m Dr. H . I,. Hansen. (4; Liacher and Jordan Trim
JGURSAL
64, lG23 (1937).
( 5 ) Cook and Kreke rbid , 61, 2071 (1Y39)
